| 1 |
On 12/22/2011 05:44 PM, Tanstaafl wrote: |
| 2 |
> On 2011-12-20 12:19 PM, Nikos Chantziaras <realnc@×××××.de> wrote: |
| 3 |
>> If you allow someone to edit root owned files, you're practically giving |
| 4 |
>> him root access. |
| 5 |
> |
| 6 |
> Well, yeah, but only on those defined files... |
| 7 |
|
| 8 |
root access is global. You can't limit it. root is root, the all |
| 9 |
powerful Unix being. Period :-) |
| 10 |
|
| 11 |
|
| 12 |
>> So the fact that he doesn't know the root password is totally |
| 13 |
>> irrelevant; he doesn't even need the password anymore to gain root |
| 14 |
>> access since he already has that access. |
| 15 |
> |
| 16 |
> But he only has root access in explicitly defined, non-system, non |
| 17 |
> critical directories... |
| 18 |
|
| 19 |
Again, root can have no limits. |
| 20 |
|
| 21 |
|
| 22 |
>> So you might want to rethink the way you want to allow him to edit those |
| 23 |
>> files. |
| 24 |
> |
| 25 |
> I *want* him to be able to do whatever he wants in /var/www (and a few |
| 26 |
> other non critical directories)... |
| 27 |
|
| 28 |
Then you put the files in a special group and make them g+w, and add the |
| 29 |
affected users to that group. Then they will able to write to those |
| 30 |
files. If you want to give them write access to a whole directory, you |
| 31 |
put the directory in the group and make it g+w. This is how it's |
| 32 |
traditionally been done in Unix for ages, and it's extremely easy to set up. |
Archives